NTAM Factor Research Leader: Get Paid for the Risk You Take

Mike Hunstad, as head of quantitative strategies for Northern Trust Asset Management, spends much of his working time considering the complicated subject of factor-based investing within institutional portfolios.

According to Hunstad, many types of institutional investors, including defined contribution (DC) and defined benefit (DB) retirement plans, have started to embrace factor-focused portfolios and smart-beta strategies as a means of achieving excess returns. With the latest bouts of market volatility there has been something of a rush into the low-volatility factor, he notes, but other factors are also receiving increased client attention.

Many institutions have found some real success with factor investing and smart-beta, Hunstad suggested, but there are some emerging challenges that investors must be made aware of. In particular, he warns that institutions may be trying to implement too many factors at the same time without considering the complex interplay between sources of portfolio risk and returns.

PLANSPONSOR: I understand you are the head of quantitative strategies for Northern Trust Asset Management. Can you define that role for us and tell us about what your daily work entails? How much are you focused on the subject of factor investing among institutional investors?

Mike Hunstad:  I’m glad you asked that, because the bulk of our mandate for quantitative analysis today is centered on factor research. My team is responsible for factor-based research and product development for equities and, increasingly, within fixed income. We have grown tremendously in the space in the last several years in terms of assets, and so the conversations we have on a daily basis really have changed as well.

When I started on this work six years ago, we were still focused on basic education around this topic—what is a factor? Why does it exist and why does it offer an opportunity for pursing outperformance? That sort of thing was our bread and butter. Years later the conversation is much more sophisticated and we are focused on some of the more complicated and challenging institutional implementation issues.

We have started to see the trend of virtually all of our clients, and all the conversations we are having with prospects, moving away from single factor implementation into a multi-factor context. This has brought really a whole new level of required sophistication and new considerations about what the potential benefits and pitfalls of factor investing are for institutional clients.

PS: I understand that in a recent review of more than 300 institutional portfolios, Northern Trust Asset Management has found a very common problem leading to underperformance. You refer to it as ‘dilution,’ and warn that a lack of sophistication around factor-based portfolio design can cause the benefits of individual factors to cancel each other out, especially in this context of institutions trying to take advantage of multiple factors at the same time. Can you give us an overview of this issue?

MH: The motivation behind this research was addressing what to some investors appears to be a contradiction. On the one hand we have spoken at length in the past about the efficiency and excess returns that can be found in factor implementation, but this efficiency has proven to be difficult for some investors to achieve in practice. The main challenge they face is that any time you tilt with real conviction towards a given factor, it doesn’t matter which, if you are not careful there is going to be a strong tendency to take unintended bets and thus to carry unintended risks. This can potentially cancel out any excess return you are going to earn from managing towards that factor.  

Consider an example. Let’s say that you have tilted towards the low-volatility factor in the last couple of months, which would make sense for many institutions in this uncertain environment. If you were not careful and thoughtful, this would have brought you towards a strong exposure to utilities and consumer staples and would probably have resulted in a large-cap equity bias, and you may even have moved towards a negative value-factor bias, meaning your stocks are more expensive on average. Overall this portfolio may serve some of your needs, but you can also very quickly start to develop a lot of unexpected, idiosyncratic risk that sneaks in, such that you think you have low volatility but in fact you have added other unanticipated risks and canceled this out. We call this ‘dilution’ of risk-adjusted returns one of the main enemies of achieving the factor performance you expect.

You can see just by looking back at the market performance how these unanticipated risks could damage portfolio performance. Anyone who took on a great deal of exposure to public utilities and real estate—which are bond proxies—in the first part of this year has not had a great time. So we see a lot of very unhappy low volatility investors sitting out there right now who found themselves being surprised by unintended risk factors they took on while attempting to craft a low-volatility strategy. The important takeaway here is that factor-based portfolios must be reviewed to ensure they are not taking unintended bets and risks. You want as pure implementation of that factor as you can get. In our low-volatility strategy, for example, we work to make sure there is not an unexpected over-concentration to utilities or indeed to any one sector. This has paid off year-to-date, because we are sector neutral, region neutral, and style-factor neutral.

PS: You also have framed the ‘dilution’ challenge in terms of ‘over-diversification.’ Can you explain what you mean by this?

MH: This challenge of getting the most out-of-factor-based portfolios, of course, also ties into the idea of diversification, and now we are increasingly talking about the concept of over-diversification at a portfolio and strategy level. If you become aware of the unintended risks in your factor portfolio but are not very careful about how you diversify away from those risks, you can also run into problems. So in the current example of moving towards the low-volatility factor, you will have to ask, how am I going to diversify away from the utilities risk, the consumer staples risk, and the other idiosyncratic risks? In answering this question, it is actually quite easy to jeopardize your factor convictions and throw the baby out with the bath water, so to speak. This is the concern we are talking about when we mention over-diversification and ‘dilution.’

Remember, the reason why we invest in factor portfolios in the first place is that we are looking to improve risk-adjusted returns. Factors have been demonstrated to both improve risk-adjusted returns against the market and also offer lower correlations. But a lot can happen between acknowledging a potential factor advantage and then putting it into practice. Whether targeting one factor, two factors or three factors, the more complicated our thesis might be, the more likely it can become that we are introducing dilution into the portfolio—especially the aggregate portfolio of a large institutional investor. In short, dilution can undue both our performance convictions and the diversification effort as we are trying to push our portfolio out onto the cusp of the efficient frontier.

The topic of dilution, viewed from both of these perspectives, is increasingly important right now, because more and more investors are seeking ways to use more than one of these factors at the same time to pursue excess returns.

PS: I have heard you use the phrase, ‘get paid for the risk you take.’ Can you talk us through in more detail what you mean by getting paid for taking risk? That’s the ultimate source of market returns, after all, so perhaps that is a good framework for understanding this whole conversation?

The number of factor-based strategies has significantly expanded in recent years as investors seek to capture excess returns from well-defined compensated risk factors such as size, value and low-volatility. Despite targeting the same factors, however, these strategies can produce very different returns. Our research has shown that this is the result of unintended exposures to uncompensated risk factors, which can range from sector concentration, unanticipated leverage, etc., that contribute to risk but not to returns.

To measure the overall appropriateness of a factor strategy—to make sure you get paid for the risk you take—we have created a new metric called the factor efficiency ratio (FER). For a strategy to have a high factor efficiency ratio, it must have a combination of high intended factor exposure and low unintended risk exposure. We confirmed this concept by analyzing the FER, to determine factor purity, and Sharpe ratios, to calculate risk-adjusted returns, of a number of popular smart beta strategies. We saw that a higher FER did in fact produce higher risk-adjusted returns.

PS: Can we take a step back and discuss the widely held belief that the ‘value factor’ has underperformed over the last decade? As a result, we hear about plan sponsors and advisers shying away from using value as an investing theme. You have suggested this is a ‘complete misperception’ and that, in reality, value has outperformed and has been the second best factor performer behind momentum.

MH: Yes, this is an important and related point to what we have covered so far. Whenever we look at factor returns, we have to be very careful about defining the return of the factor itself versus defining the return of the particular implementation of the factor.

Let’s say you have a value indexed investment right now, and you look at this index and it shows you have underperformed the broader market, do you jump to the conclusion that the value factor is to blame? Using the same logic we have been discussing, you can’t really do that. When you look at the value index, it may actually be that your particular implementation of value involved some sector-concentration bets that did not end up paying off—or you might have had other idiosyncratic risks or style factor exposures that did not pay off.

The lesson is that, we cannot be so quick to say that ‘value has underperformed’ or ‘low-volatility has underperformed,’ without first going in and doing the quantitative work of disaggregating the returns in terms of the influence of the factor or factors, versus the influence of those unintended bets and exposures that have crept in along with your factor exposure.

So, sticking with value, we have seen clear evidence that you have to be really careful about making a claim about the performance of any given factor without taking into account all the other elements of risk and return. When you do this and back out some of the sector bets inherent in value indexes, the picture changes dramatically. We see that the influence of these unintended bets can be really dramatic and totally outshine the influence of the factor in question.